PJ Media

“We don’t do body counts” – US General Tommy Franks

As Sum Guy pointed out above,

1) The study provides estimates of increased mortality, not just military deaths

2) There’s a 95% chance that if they did the study again with the same biases, their estimates could be as “low” as 300,000.

My comments:

1) Passive body-count methods are unlikely to capture the extent of “regular” mortality.

2) The cluster sampling methodology is very common, has been used by the WHO and others to provide health-related estimates

3) Because deaths are so rare (5.5/1000), cluster sampling is less accurate

4) The binomial distribution for variance estimation might be more appropriate than the normal distribution, given the rarity of deaths

5) The fact that their recent study generated estimates nearly the same as the 2004 survey suggests that the methodology is reliable, if not entirely accurate

6) Assuming that the bias in the recent study is the same as the 2004 survey, there has clearly been a large increase in mortality in Iraq

I don’t see anything fundamentally wrong with their methodology. It’s safe to say that the increase in mortality is at least 300,000. Regardless of the number of Iraqis who died because they were directly involved in conflict with the CPA or other Iraqis, it’s a terrible loss of life. The US is partly responsible.